Metastasis remains the main cause of death in cancer patients, and current treatment options are insufficient. There is a great need to define new strategies. The process of metastasis is regulated by specific sets of metastasis genes. Whereas some metastasis genes may be common for all cancers, some may be specific for distinct cancer types. Characterization of such genes holds great promise for the future because they may serve as targets for disease monitoring and therapy. In the past, two parallel but individually incomplete genomic approaches were established to define such genes, the in-vivo selection of derivative cell lines with high metastatic potential and gene expression profiling of human tumor tissue. However, each approach is associated with pitfalls, a reason that may account for the relatively poor overlap of findings among different groups. Therefore, rigorous validation of the biological relevance of such metastasis genes for distinct cancer types is of the utmost importance.